Re: [Boost-users] [Serialization] [Portable Binary]

Subject: Re: [Boost-users] [Serialization] [Portable Binary]
From: William Oquendo (woquendo_at_[hidden])
Date: 2010-09-12 12:35:21

Hi, thanks for the reply.

In the final version of the program, I needed to manipulate also float
values. I was testing on Snow Leopard + gcc 4.5 , then, when running the
same code with the same binary data, the program did not run both on Lunix
Slackware64 13.1 + gcc 4.4.4,  and  Linux Slackware 23.0 + gcc 4.4.4 (and
gcc 4.3). SO, temporarily, I had to rewrote everything to handle only text
files, and now everything is working.

I will try to isolate the issue in a smaller code to check if this is a
compiler issue, a platform issue, or a boost::binary issue.

Best regards / Cordialmente,

--
William Oquendo
Phd Candidate
Simulation Of Physical Systems Group
Universidad Nacional de Colombia
Linux User # 321481
*********************
Este correo puede carecer de tildes o eñes ya que el teclado no contiene
estos caracteres. Presento excusas por eso.
*********************
On Sat, Sep 11, 2010 at 12:21 AM, Robert Ramey <ramey_at_[hidden]> wrote:
>  William Oquendo wrote:
> > Dear,
> >
> > In the documentation of boost::serialization, I read that there is
> > still work needed to have a portable (32 and 64 bits) binary
> > serialization implementation. The date of the TODO is 2008. Is still
> > this a current TODO? is there a library implementing this? currently,
> > It is not a matter of dead or live, but I would like to know. I have
> > been trying to run a code in different (bits) systems, and I got the
> > error "Incompatible native format - size of int"
>
> This should only occur when trying to load data which
> cannot be represented in the loading system.  For example
> suppose one has a value >32 bits on a 64 bit machine.  No
> problem. Then it is saved to a portable binary archive.  Still
> no problem.  Then it is loaded by a mache with 32 bit integers.
> Uh-oh - the information can't be represented and an
> exception is thrown.
>
> I believe that the only think pending is a more general
> representation of floating point values.  Apparently that
> hasn't been an issue for people that use it.
>
> Robert Ramey
>
> >
> > Best regards / Cordialmente,
>
> _______________________________________________
> Boost-users mailing list
> Boost-users_at_[hidden]
> http://lists.boost.org/mailman/listinfo.cgi/boost-users
>